The marRAB multiple antibiotic resistance operon of Escherichia coli controls the expression of a large number of genes resulting in low level antibiotic and superoxide resistance through a complex network of reactions. MarR auto-represses the mar operon but is inactivated upon interaction with salicylate, losing its DNA binding capacity. This, in turn, results in derepression of the operon and expression of MarA, which activates the transcription of some 40 to 60 promoters (the mar/sox/rob regulon) including the marRAB promoter itself (auto-activation). MarA is one of 22 members of the AraC family of transcriptional activators in Escherichia coli. For a large subclass of the promoters activated by AraC family members, namely those where the DNA binding region overlaps the DNA binding region of RNA polymerase (Class II promoters), a polar interaction between the sigma subunit of RNA polymerase and the activator is thought to be essential. Two members of the AraC family, SoxS and Rob, are very closely related to MarA. So much so, that although they are separately controlled they activate the same 40 to 60 promoters as MarA. MarA shares >90% homology with SoxS in that portion of the molecule essential for binding to the DNA consensus sequence. It is 100% homologous with SoxS in that portion of the molecule that interacts with the alpha subunit of RNA polymerase. Yet, it shares no homology whatever with that region of SoxS that has been reported to be responsible for the polar interaction with the sigma subunit. By a variety of techniques we have now shown that MarA also requires polar interactions with the sigma subunit at one of these Class II promoters, tolC, and that the site of this interaction is different in MarA and SoxS. Furthermore, preliminary evidence suggests that MarA utilizes different negative surface charges for interaction with sigma at different Class II promoters helping to explain its high degree of versatility. We have also focused recently on other aspects of the differences in activation by MarA and SoxS and a mathematical model is being developed in an effort to integrate these aspects into the overall model for activation. The remarkable conclusion of this modeling is that the activation of most genes of the regulon is best accounted for by a very large increase in the forward reaction rate (clearance) accompanied by a decrease (rather than an increase) in the affinity of RNA polymerase for the promoter in the presence of the activator. [unreadable] This work was carried out in collaboration principally with Drs. J.L. Rosner and Michael Wall (Computer and Computational Sciences & Bioscience, Los Alamos National Laboratory, Mail Stop B256,Los Alamos NM 87545 USA).